Saturday, December 3, 2022
Home TECH The challenges of verifying AI for healthcare

The challenges of verifying AI for healthcare

There is a lot of excitement in healthcare about the use of artificial intelligence (AI) to improve clinical decision making.

Pioneered by IBM Watson for Healthcare and DeepMinds Healthcare, AI promises to help specialists diagnose patients more accurately. Two years ago, McKinsey co-produced a report with the European Union EIT Health to explore the potential of AI in healthcare. Among the key opportunities the report authors found were in healthcare operations: diagnostics, clinical decision support, triage and diagnosis, care delivery, chronic care management, and self-care.

“First, the solutions are likely to address the routine, repetitive, and largely administrative tasks that absorb significant physician and nurse time, streamlining healthcare operations and increasing adoption,” they wrote. “In this first phase we would also include image-based AI applications, which are already in use in specialties such as radiology, pathology and ophthalmology.”

The world of AI in healthcare has not stopped and in June, the European Parliament published artificial intelligence in health focusing on applications, risks, ethical and social impacts. The authors of the article recommended that AI risk assessment should be domain-specific, because clinical and ethical risks vary in different medical fields, such as radiology or pediatrics.

The authors of the article wrote: “In the future regulatory framework, validation of medical AI technologies should be harmonized and strengthened to assess and identify multifaceted risks and limitations by evaluating not only model accuracy and robustness, but also algorithmic fairness, clinical safety, clinical acceptance, transparency and traceability.”

The validation of medical AI technologies is the key focus of research carried out by the Erasmus University Medical Center in Rotterdam. Earlier this month, Erasmus MC, University Medical Center Rotterdam, began working with health technology company to launch its AI Innovation Lab for Medical Imaging.

The initial program will run for three years and will carry out detailed research on abnormality detection using AI algorithms for infectious and non-infectious diseases. The researchers hope to understand the potential use cases for AI in Europe and provide guidance to clinicians on best practices for adopting the technology specifically for their requirements.

Jacob Visser, Radiologist, Chief Medical Information Officer (CMIO) and Assistant Professor of Values-Based Imaging at Erasmus MC, said: “It’s important to realize that we have big challenges, an aging population, and we have a lot of technology that needs to be used effectively. responsibly. We are investigating how we can bring value to doctors and patients using new technologies and how we can measure those advances.”

Visser’s role as CMIO acts as a bridge between the medical side and the technologists. “As a medical professional, the CMIO wants to steer IT in the right direction,” she said. “Doctors are interested in the possibilities offered by IT. New technical developments make doctors see greater opportunities in areas such as precision medicine.”

Erasmus MC will run the lab and conduct research projects using Qure’s AI technology. The initial research project will focus on musculoskeletal and thoracic imaging. Visser said that by evaluating the AI ​​models, “it’s easy to verify that a fracture has been correctly detected.”

This makes it possible to assess how well the AI ​​is coping, allowing researchers to gain meaningful insight into how often the AI ​​incorrectly misses a genuine fracture (false negative) or misclassifies an X-ray scan. as a fracture (false positive).

Speaking about the level of scrutiny that goes into the use of AI in healthcare, Vissier said: “Medical algorithms need to be approved, for example, by the Federal Drug Administration [FDA] in the US and get CE certification in Europe.”

Regarding the partnership with, he added: “We see the adoption of AI in healthcare at a critical time, where clinicians are asking for expert advice on how best to assess the adoption of the technology. In Qure’s work to date, it is clear that they have collected detailed information on the effectiveness of AI in healthcare settings, and together we will be able to assess effective use cases in European clinical settings.”

But there are many challenges in using AI for healthcare diagnoses. Even if an algorithm has been approved by the FDA or has CE certification, this does not necessarily mean that it will work in a local clinical practice, Vissier said. “We have to make sure the AI ​​algorithm meets our local practice needs,” she added. “What are the clinically relevant parameters that can be affected by the results that AI produces?”

The challenge is that the data used to develop an AI algorithm for healthcare uses a specific data set. As a consequence, the resulting data model may not be representative of actual patient data in the local community. “You see a performance drop when you validate an algorithm externally,” Vissier said.

This is analogous to pharmaceutical trials, where side effects can vary between populations. The pharmaceutical industry monitors usage, which feeds into the product development cycle.

As for his aspirations for research coming out of the new lab, Vissier said: “I hope, within a year, to prove that the algorithms work, the accuracy of their diagnoses, and I hope that we have started to evaluate how these algorithms work in practice.” daily clinic.


How a small electoral business became a conspiracy theory target

At an invitation-only conference in August at a secret location southeast of Phoenix, a group of election deniers revealed a new conspiracy theory about...

19 Android Settings You Might Not Know About

if you are a typical smartphone user, you average almost five hours a day on your phone now. But beyond video and social...


Please enter your comment!
Please enter your name here

- Advertisment -

Most Popular

Time to get serious about Democrats’ chances in the House

There were more than a few Democrats who were a little miffed at my Friday bulletin on election rigging, which argued that Democrats aren't...

Uncertain Path For US-Taiwan Free Trade Deal Despite Hill’s Support

If Taiwan's semiconductor industry were to be destroyed, downgraded or subjected to Western sanctions as a result of a Chinese military occupation of the...

As Ukrainian forces enter Lyman, hundreds of Russians may still be trapped inside

A local resident walks past abandoned Russian tanks in recently liberated Kyrylivka. September 30, 2022. All through Friday night, Russian sources on Telegram were...